The presentation and course of schizophrenia have implicated brain dysfunction. Earlier work was limited by lack of tools for assessment of the nature of potential aberrations in brain processes. Development of neuroimaging techniques have provided an opportunity for measuring brain function in search for neural substrates of schizophrenia. The proposed study will use 15O-H2O to measure regional cerebral blood flow (CBF) in a repeated-measures design to evaluate the regional specificity of abnormalities in brain activation response to cognitive challenge tasks. It is also proposed to relate the pattern of resting glucose metabolism to course of illness. Patients with a range of chronicity will be followed longitudinally on a standard pharmacologic intervention. PET FDG studies will be obtained repeatedly and glucose metabolism will be related to clinical course and CBF. Finally it is proposed to examine the relationship between abnormalities of regional brain activity and clinical, neuropsychological and neuroanatomic measures. It is hypothesized that abnormalities in the magnitude of regional increase occur with stimulation in addition to abnormalities of resting CBF, and the former are more revealing of the neural substrates of symptoms. Each of the proposed experiments will probe brain regions implicated in specific tasks. Predications: a. Memory tasks produce most pronounced activation in temporal lobe regions in normals, left higher for verbal memory and right for facial memory. Schizophrenia is associated with failure to activate, more evidenced in left hemisphere. b. Abstraction is associated with frontal lobe activation in normals and failure of frontal lobe activation in patients. c. The abnormality of frontal lobe activation for abstraction is less pronounced than that for memory. d. Impaired semantic classification and syntax are associated with the right hemispheric abnormalities. e. Attentional dysfunction will be associated with abnormalities of activation in frontal and parietal regions. f. Regional abnormalities in activation will be more highly correlated than resting flow patterns with symptom severity and subtype.